From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-bugzilla@gcc.gnu.org>
Received: by sourceware.org (Postfix, from userid 48)
	id D2A233858C27; Thu, 14 Sep 2023 06:46:13 +0000 (GMT)
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org D2A233858C27
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org;
	s=default; t=1694673973;
	bh=zgY3R03e5bYGzg74EWq4rGq4BWpe2b6sk+JVbr8eylI=;
	h=From:To:Subject:Date:In-Reply-To:References:From;
	b=B+kOT2kOEnszN1fPgBrVYkLrpiq0lGYqsbrRkNwDtFFwfja3Zqa1aX5s0w/A9Nutc
	 HG3SJ3cWaNVYKaIO3vNyOR1agYu34QsCehedvtM2gA/ggV2hprD/qrvS9IjD1zmcTv
	 9KE8ZRf1Gmh2gRYZQeg6winwdD25kgvIJdRlvlMQ=
From: "rguenther at suse dot de" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug middle-end/111401] Middle-end: Missed optimization of
 MASK_LEN_FOLD_LEFT_PLUS
Date: Thu, 14 Sep 2023 06:46:12 +0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: middle-end
X-Bugzilla-Version: 14.0
X-Bugzilla-Keywords: missed-optimization
X-Bugzilla-Severity: normal
X-Bugzilla-Who: rguenther at suse dot de
X-Bugzilla-Status: NEW
X-Bugzilla-Resolution: 
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Flags: 
X-Bugzilla-Changed-Fields: 
Message-ID: <bug-111401-4-kiStXFIV6e@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-111401-4@http.gcc.gnu.org/bugzilla/>
References: <bug-111401-4@http.gcc.gnu.org/bugzilla/>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/
Auto-Submitted: auto-generated
MIME-Version: 1.0
List-Id: <gcc-bugs.sourceware.org>

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=3D111401

--- Comment #4 from rguenther at suse dot de <rguenther at suse dot de> ---
On Wed, 13 Sep 2023, rdapp at gcc dot gnu.org wrote:

> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=3D111401
>=20
> Robin Dapp <rdapp at gcc dot gnu.org> changed:
>=20
>            What    |Removed                     |Added
> -------------------------------------------------------------------------=
---
>                  CC|                            |rdapp at gcc dot gnu.org
>=20
> --- Comment #2 from Robin Dapp <rdapp at gcc dot gnu.org> ---
> I played around with this a bit.  Emitting a COND_LEN in if-convert is ea=
sy:
>=20
> _ifc__35 =3D .COND_ADD (_23, init_20, _8, init_20);
>=20
> However, during reduction handling we rely on the reduction being a gimple
> assign and binary operation, though so I needed to fix some places and in=
dices
> as well as the proper mask.
>=20
> What complicates things a bit is that we assume that "init_20" (i.e. the
> reduction def) occurs once when we have it twice in the COND_ADD.  I just
> special cased that for now.  Is this the proper thing to do?

I think so - we should ignore a use in the else value when the other
use is in that same stmt.

> diff --git a/gcc/tree-vect-loop.cc b/gcc/tree-vect-loop.cc
> index 23c6e8259e7..e99add3cf16 100644
> --- a/gcc/tree-vect-loop.cc
> +++ b/gcc/tree-vect-loop.cc
> @@ -3672,7 +3672,7 @@ vect_analyze_loop (class loop *loop, vec_info_shared
> *shared)
>  static bool
>  fold_left_reduction_fn (code_helper code, internal_fn *reduc_fn)
>  {
> -  if (code =3D=3D PLUS_EXPR)
> +  if (code =3D=3D PLUS_EXPR || code =3D=3D IFN_COND_ADD)
>      {
>        *reduc_fn =3D IFN_FOLD_LEFT_PLUS;
>        return true;
> @@ -4106,8 +4106,11 @@ vect_is_simple_reduction (loop_vec_info loop_info,
> stmt_vec_info phi_info,
>            return NULL;
>          }
>=20
> -      nphi_def_loop_uses++;
> -      phi_use_stmt =3D use_stmt;
> +      if (use_stmt !=3D phi_use_stmt)
> +       {
> +         nphi_def_loop_uses++;
> +         phi_use_stmt =3D use_stmt;
> +       }
>=20
> @@ -7440,6 +7457,9 @@ vectorizable_reduction (loop_vec_info loop_vinfo,
>        if (i =3D=3D STMT_VINFO_REDUC_IDX (stmt_info))
>         continue;
>=20
> +      if (op.ops[i] =3D=3D op.ops[STMT_VINFO_REDUC_IDX (stmt_info)])
> +       continue;
> +
>=20
> Apart from that I think what's mainly missing is making the added code ni=
cer.=20
> Going to attach a tentative patch later.
>=20
>=